From mboxrd@z Thu Jan 1 00:00:00 1970 From: sougata santra Subject: Re: [PATCH 1/1] hfsplus: skip unnecessary volume header sync Date: Sat, 19 Jul 2014 14:59:00 +0300 Message-ID: <53CA5D84.5090309@tuxera.com> References: <1405614762.25052.8.camel@ultrabook> <1405672120.2626.32.camel@slavad-CELSIUS-H720> <1405675453.25052.36.camel@ultrabook> <3E310BE8-6D1D-439D-9E98-E2C83270D1F1@dubeyko.com> Mime-Version: 1.0 Content-Type: text/plain; charset="ISO-8859-1"; format=flowed Content-Transfer-Encoding: 7bit Cc: Andrew Morton , , , Fabian Frederick To: Vyacheslav Dubeyko Return-path: Received: from nbl-ex10-fe01.nebula.fi ([188.117.32.121]:54671 "EHLO ex10.nebula.fi" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1754528AbaGSL7E (ORCPT ); Sat, 19 Jul 2014 07:59:04 -0400 In-Reply-To: <3E310BE8-6D1D-439D-9E98-E2C83270D1F1@dubeyko.com> Sender: linux-fsdevel-owner@vger.kernel.org List-ID: On 07/19/2014 02:23 PM, Vyacheslav Dubeyko wrote: > > On Jul 18, 2014, at 1:24 PM, Sougata Santra wrote: > >> On Fri, 2014-07-18 at 12:28 +0400, Vyacheslav Dubeyko wrote: >>> On Thu, 2014-07-17 at 19:32 +0300, Sougata Santra wrote: >>>> hfsplus_sync_fs always updates volume header information to disk with every >>>> sync. This causes problem for systems trying to monitor disk activity to >>>> switch them to low power state. Also hfsplus_sync_fs is explicitly called >>>> from mount/unmount, which is unnecessary. During mount/unmount we only want >>>> to set/clear volume dirty sate. >>>> >>> >>> As far as I can judge, hfsplus driver has hfsplus_mark_mdb_dirty() >>> method. This method "marks" volume header as dirty and to define some >>> dirty_writeback_interval. The hfsplus_mark_mdb_dirty() is called in such >>> important methods as hfsplus_block_allocate(), hfsplus_block_free(), >>> hfsplus_system_write_inode(), hfsplus_link(), hfsplus_new_inode(), >>> hfsplus_delete_inode(). So, it means for me that every call of >>> hfsplus_sync_fs() is made when volume header should be written on >>> volume. So, if you can detect some inefficiency or frequent calls of >>> hfsplus_sync_fs() then, maybe, it needs to optimize >>> hfsplus_mark_mdb_dirty() in the direction of proper >>> dirty_writeback_interval definition. What do you think? >> >> Thanks a lot for taking time to look into the patch. I will look into >> it. hfsplus_sync_fs() also called explicitly from mount/unmount (It is >> not called from remount, that is a bug and needs to be addressed. ). >> This is not required at all since it is already called from vfs. The >> only purpose of calling them from mount/unmount is to update dirty/ >> clear state and other info like driver version, write count etc ... >> When clearly hfsplus_sync_fs() does more than updating volume header >> and flushing it to disk. >> > > Finally, I think that more compact and better solution can be achieved by > modification of hfsplus_mark_mdb_dirty() using. > > We can add into struct hfsplus_sb_info two additional fields: > (1) flag that informs about dirty state of volume header; > (2) last write timestamp. > > So, every calling of hfsplus_mark_mdb_dirty() will set flag of volume header > dirtiness. The hfsplus_sync_fs() will know that volume header is really dirty > on basis of checking dirty flag in hfsplus_sb_info. And hfsplus_sync_fs() will > clear this flag after saving state of volume header and special files on volume. > > The last write timestamp + timeout between two adjacent writes can be used > for decreasing frequency of flushing of volume header on volume. Such technique > is used in NILFS2. Thank you for the pointer, will look into it. I guess it has to be locked against mutual access as well. > > [snip] >>> >>>> if (!error) >>>> error = error2; >>>> } >>>> @@ -206,34 +256,41 @@ static int hfsplus_sync_fs(struct super_block *sb, int wait) >>>> >>>> mutex_lock(&sbi->vh_mutex); >>>> mutex_lock(&sbi->alloc_mutex); >>>> - vhdr->free_blocks = cpu_to_be32(sbi->free_blocks); >>>> - vhdr->next_cnid = cpu_to_be32(sbi->next_cnid); >>>> - vhdr->folder_count = cpu_to_be32(sbi->folder_count); >>>> - vhdr->file_count = cpu_to_be32(sbi->file_count); >>>> >>>> - if (test_and_clear_bit(HFSPLUS_SB_WRITEBACKUP, &sbi->flags)) { >>>> - memcpy(sbi->s_backup_vhdr, sbi->s_vhdr, sizeof(*sbi->s_vhdr)); >>>> - write_backup = 1; >>>> + free_blocks = cpu_to_be32(sbi->free_blocks); >>>> + next_cnid = cpu_to_be32(sbi->next_cnid); >>>> + folder_count = cpu_to_be32(sbi->folder_count); >>>> + file_count = cpu_to_be32(sbi->file_count); >>>> + >>>> + /* Check if some attribute of volume header has changed. */ >>>> + if (vhdr->free_blocks != free_blocks || >>>> + vhdr->next_cnid != next_cnid || >>>> + vhdr->folder_count != folder_count || >>>> + vhdr->file_count != file_count) { >>> >>> I don't think that this check is correct because volume header contains >>> some flags and forks. >> >> Can you please elaborate ? What are the other forks and flags that gets >> updated in volume header. >> > > There is simple use-case. For example, you mount file system, do nothing > with it and, finally, unmount file system. You should set > kHFSBootVolumeInconsistentBit in attributes field of volume header during > mount. So, you should write new state of volume header on volume. > Then, during unmount, you should set kHFSVolumeUnmountedBit in > attributes field of volume header and save this state of volume header on > volume. This flags are used by fsck and file system driver for detecting > presence of error on volume. > Please look at the patch, it separates out the mount/unmount code flow from sync file system code flow. Volume header is modified by both when syncing file-system (because changes in allocations..etc), and during mount and unmount. These are separate context and should not be mixed with each other, please see any other file-system. When volume dirty/clean states are updated, other version info etc are updated they should be immediately flushed to disk that is it, but sync file system context is much more than that. AFAIK the relevant checks in sync_fs does some check to see if any allocation states have changed, I agree this can be done in hfsplus_mark_mdb_dirty but I need to check that. Although the necessary checks that are done are only checking if the volume header needs to be updated and to the best of my knowledge it checks all the relevant fields which can get modified from hfsplus_sync_fs() context (not mount/unmount context). >>> Moreover, there is specially dedicated method for >>> "marking" volume header as dirty (I mean hfsplus_mark_mdb_dirty() >>> method). So, I don't think that this check is really necessary. And, >>> moreover, I don't think such significant modification of >>> hfsplus_sync_fs() makes sense at all. >> > > [snip] >>> >>>> >>>> return error; >>>> @@ -287,7 +344,7 @@ static void hfsplus_put_super(struct super_block *sb) >>>> vhdr->attributes |= cpu_to_be32(HFSPLUS_VOL_UNMNT); >>>> vhdr->attributes &= cpu_to_be32(~HFSPLUS_VOL_INCNSTNT); >>>> >>>> - hfsplus_sync_fs(sb, 1); >>>> + hfsplus_sync_volume_header(sb); >>> >>> I doubt that to flush the volume header only is proper approach. Could >>> you guarantee that special metadata files have been flushed before? >> >> Please see the cover-letter. I think that hfsplus_sync_fs is already >> called from vfs. > > Anyway, I prefer to leave code here unchanged. It is really important to guarantee > that we will have file system in consistent way after unmount. If special files were > flushed previously then it will done nothing here for special files. I don't really understand. > > Thanks, > Vyacheslav Dubeyko. > Thanks Sougata.